5/4/2010 - Question 1.  I have been hearing more frequently the comment that beta cell failure is inevitable in the TypeII diabetic. At what point does it become inevitable?at birth?with weight gain?with age?with the use of hypoglycemic medicines?I'm not sure I agree that beta cell failure is inevitable.

Thank you for this excellent question; the answer is far from straightforward. This statement, like many others floating out there, is based in some truth, but is also one that is easily misinterpreted. The concept really arose from the long follow-up experience of the UKPDS (Diabetes 1995; 44: 1249-58; and Lancet 1998; 352: 837-53) which demonstrated that there is a gradual loss of glycemic control that occurs at the same rate over time, irrespective of whether you use an insulin-sparing agent like metformin, an agent that increases beta cell workload like sulfonylureas, or even exogenous insulin which supposedly bypasses the beta cell altogether; it appeared to be part of the natural history of the disease, regardless of how you bring glucose levels down, and occurred even in spite of sustained therapy and reduction of microvascular complications with intensive control. Now, we know that the progressive decline in beta cell function occurs along the same continuum as the development of diabetes in the first place, so I think it’s important to keep this concept of “failure” in perspective: In the earlier stages of the disease, beta cell “failure” relates to the inability to meet the metabolic demands of maintaining normoglycemia imposed by peripheral insulin resistance, even though beta cell secretory capacity is often still functioning above normal capacity at that time. However, there is also a well-known progressive loss of total beta cell mass with longstanding type 2 diabetes (nicely shown by Butler et al: Diabetes 2003; 52: 102-10), so “failure” can also be due to reduced viability/survival of functional beta cells. Thus, it’s important to distinguish which aspect of beta cell “failure” we are talking about: the minute-to-minute secretory capacity to maintain normoglycemia, or the long-term balance between beta cell regeneration and apoptosis that dictates the remaining mass of functional beta cells. These two processes do have some common underlying pathogenic processes, but we don’t yet clearly understand how extensively these two aspects are interrelated in the human model. So, the question of “inevitability” can be applied to both considerations separately, and the answers are different. There are many factors that may determine the lifelong capacity of a person’s beta cells to meet the whole array of metabolic demands that may occur throughout life (so as to avoid eventual “failure”). It’s essentially impossible to enumerate all of these factors because not only do we not yet have a complete understanding of all contributing factors (e.g., genetic polymorphisms, congenital or developmental factors), but also because they are often dynamic factors that vary throughout the lifespan, including some that you have identified in your question: age, weight, medications; factors that change both on a short-term time scale (e.g., medications) as well as a long-term time scale (e.g., aging). And to make matters more complicated, almost all of what we currently understand about the effect of these factors on beta cell biology comes from in vitro studies or experimental animal models rather than human specimens, for the simple reason that animals can be sacrificed to directly examine islets, while there is no direct, non-invasive method to accurately examine human islets or quantitate regeneration, apoptosis or functional beta cell mass in free-living humans. We can study global beta cell function in humans using dynamic, stimulation testing to determine short-term secretory capacity, but we cannot currently visualize the beta cells in the in vivo human model to determine long-term cellular viability. The major exogenous insults that can occur throughout life to a type 2 patient’s beta cells are believed to include a) chronic free fatty acid elevation (“lipotoxicity”), b) chronic or cumulative elevations of glucose itself (“glucotoxicity”), c) accumulation of intra-islet amyloid deposits which appear to be dependent on the first two factors, and d) chronic inflammatory insults which may also be a manifestation of “glucolipotoxicity”. Insulin resistance, especially in conjunction with central obesity, is typically associated with elevated free fatty acid levels from unsuppressed lipolysis. While free fatty acids actually enhance beta cell insulin secretion in the short-term, long-term exposure can be detrimental and lead to a reduced insulin secretory capacity, production of reactive intermediates, and as a consequence, early apoptosis in experimental models, leading to reduced beta cell survival. A similar phenomenon occurs with chronically elevated glucose exposure: while short-term hyperglycemia obviously stimulates normal insulin secretion, long-term or repeated hyperglycemia in experimental models can lead to intracellular reactive intermediates that not only impair insulin secretory capacity, but can also lead to earlier beta cell apoptosis. Islet amyloid polypeptide is normally co-secreted with insulin from beta cells, but in the presence of metabolic insults like elevated glucose or fatty acids, can gradually lead to deposition of amyloid aggregates in the islets that is closely linked to the loss of beta cell mass and reduced functional capacity. Severe inflammation is known to impair beta cell function through the production of reactive oxygen species, but lower-grade inflammation may also represent a common pathway by which “glucolipotoxicity” leads to beta cell demise. (For reviews of these complex processes, see Pitout and Robertson: Endocr Rev 2008; 29: 351-66; and Hull et al: J Clin Endocrinol Metab 2004; 89: 3629-43). Thus, each of these acquired insults has the potential to damage beta cells both in the short- and long-term. In the intact human model, we can measure how insulin secretory capacity is adversely affected by each, but since we cannot visualize the long-term changes in beta cell morphology in free-living humans, we do not yet know for certain that each of these factors directly contributes to long-term dropout of beta cells. So, let me paraphrase the different parts of your question separately: 1) The timecourse, when it starts (birth?), and its relationship to age? When it “starts” is a very intriguing question, and not well understood. The current epidemic of early type 2 diabetes in obese youths demonstrates that secretory “failure” can certainly begin in early childhood if the genetic background and adverse environmental influences are severe enough. In theory, beta cells could be susceptible to metabolic insults anytime after the islets are formed. The well-known association between low birth weight neonates and diabetes later in life (see the review by Hill and Duvillie: Pediatr Res 2000; 48: 269-74) suggests that any inborn limitation of beta cell mass due to fetal under-development may very well contribute to an earlier diabetes onset in adulthood, which is evidence that beta cell mass as determined either genetically or developmentally may in fact be a very important factor for long-term beta cell viability. In this context, risk for “inevitable” future beta cell “failure” may have been pre-programmed even before birth. Aging, of course, affects tissues in a multitude of ways at the cellular and molecular level that we still don't completely understand. Diabetes incidence correlates directly with increasing age, but aging is also closely correlated with weight gain and changes in body composition (e.g., increasing fat mass, decreasing lean mass), which actually accounts for much of the burden of metabolic diseases seen with aging, although not all of it (see the review by Muller et al: Semin Nephrol 1996; 16: 289-98). “Lipotoxicity” becomes more severe with the central adiposity that typically occurs with aging, and “glucotoxicity” can accumulate over a lifetime as a non-diabetic or “pre-diabetic” individual ages. So, it’s not at all clear if “aging” of the beta cells per se is an independent factor in amongst all of the metabolic insults that come with aging. Regardless, there does not seem to be an age-threshold effect to the “inevitability”; it’s probably all a continuous function of the degree of exposure to these insults over time. 2) Association with weight gain? For most (but not all) patients, weight gain represents a worsening of insulin resistance (and therefore an increased beta cell workload to try to maintain normoglycemia, which in turn may predispose to more “glucotoxicity”), but it also usually comes with greater “lipotoxicity” and greater severity of inflammation. There is also substantial heterogeneity in the population: We’ve all met patients who, despite being quite obese, do not manifest any of the classic metabolic features of insulin resistance, while others who are perhaps not as obese may manifest much more advanced metabolic derangements. The same degree of weight gain may have different adverse consequences for different people, especially across racial groups, so differences in body fat distribution (visceral vs. subcutaneous) are also another important variable that determines the severity of adverse metabolic consequences. Thus, beta cell damage is clearly not simply a matter of the degree of obesity, but more likely the myriad of factors for which weight is simply a marker (not to mention the varied genetic background). The benefits of weight loss in preventing new diabetes (e.g., XENDOS: Diabetes Care 2004; 27: 155-61; SOS: N Eng J Med 2004; 351: 2683-93; or the DPP: N Eng J Med 2002; 346: 393-403) would seem to suggest that beta cell “failure” due to weight gain is not “inevitable”, but these studies measured glucose endpoints in establishing new-onset diabetes, and they do not tell us whether we have actually changed the long-term survival of the functional beta cell mass. Regardless, there does not seem to be a weight-related threshold for “inevitability”; it’s probably all a continuum of factors that are both dynamic and heterogeneous. 3) Association with hypoglycemic medications? Of course, it totally depends on which medications you want to talk about, since they work through such diverse mechanisms. In general, drugs that act through sensitization of peripheral tissues will reduce the metabolic demand and therefore alleviate beta cell “failure” by reducing its workload and/or maintaining normoglycemia for longer. There is also evidence that sensitizers may have direct, protective effects on beta cells: Metformin improves insulin secretion and protects from “glucolipotoxicity” both in vitro and in vivo, and clearly reduces hyperglycemia as well as progression to diabetes in the DPP trial as compared to placebo; but it has not been shown to slow the progression of long-term beta cell demise in diabetic patients, since the rate of deterioration in glycemic control in the UKPDS was no different than that of other agents. Thiazolidinediones (glitazones) are potent insulin sensitizers that can also ameliorate the adverse effects of “glucolipotoxicity”, and in some experimental models, they appear to even prolong beta cell survival. The DREAM Study (Lancet 2006; 368: 1096-105) and trials in women with gestational diabetes (TRIPOD: Diabetes 2002; 51: 2796-803; PIPOD: Diabetes 2006; 55: 517-22) showed that they can reduce the incidence of new-onset diabetes, while the ADOPT Study (N Eng J Med 2006; 355: 2427-43) showed that they can sustain the efficacy of monotherapy better than other agents. However, these trials were all of relatively shorter duration (3-5 years), and could not directly examine long-term beta cell survival per se. Both ADOPT and data from the original troglitazone arm of the DPP (Diabetes 2005; 54: 1150-6) suggested that thiazolidinediones may be superior to other agents in slowing disease progression (i.e., a shallower slope of worsening glycemia / beta cell secretory decline), but it’s still unclear if this was merely a potency phenomenon or a qualitatively different beta cell effect beyond just enhancing secretion. In the longer studies (ADOPT, DREAM), there was still a gradual worsening of glycemic control with ongoing follow-up (consistent with the UKPDS experience), indicating that despite effectively slowing disease progression, they still could not halt the progression of disease completely. Conversely, it has been suggested that sulfonylureas, by virtue of being secretagogues and not sensitizers, may somehow contribute to beta cell “failure” (by “working them harder”). While this may seem like a logical theory, and is supported by some in vitro studies suggesting that secretagogues may increase beta cell apoptosis, there is as yet no clinical evidence that it actually results in beta cell “harm” in humans. If “overwork” somehow kills off beta cells faster, then the 10-year follow-up of UKPDS should have shown a more rapid worsening of glycemic control with sulfonylureas than other treatments, and that was not the case. Also, although ADOPT suggested that thiazolidinediones were superior to sulfonylureas, there was no placebo comparison group, so one cannot conclude from ADOPT that sulfonylureas were inherently detrimental, and also since short-term glycemia still improved with sulfonylureas. Estimates of beta cell function in ADOPT were initially greater with sulfonylureas than thiazolidinediones, and while its decline with time was faster, the absolute level of function achieved with sulfonylureas by the end of the 5-year follow-up was no worse than thiazolidinediones. Newer agents that act through incretin pathways (GLP-1 agonists, DPP-4 inhibitors) may show some promise in this regard, since incretins have been shown to actually promote growth and differentiation of beta cells in culture and animal models, as well as slow the rate of beta cell apoptosis that occurs with “glucolipotoxicity” in experimental model systems (see the review by Salehi et al: Endocr Rev 2008; 29: 367-79). These pre-clinical data suggest that modulation of the incretin system may represent the potential to substantially maintain or maybe even expand beta cell mass in spite of those metabolic insults. As of now, however, because there is no way to non-invasively measure beta cell mass in humans, and because there are differences in the physiology of rodent and human beta cells, whether these agents will actually enhance long-term beta cell survival in humans remains to be established. Sustained use of exenatide for over 3 years in uncontrolled, open-label follow-up of self-selected trial participants has so far demonstrated sustained benefits on glucose lowering, weight loss, and apparent preservation of estimated beta cell function, but this does not yet rival the longer, properly-designed follow-up of other major trials. 4) Is it truly “inevitable”? Oxford American Dictionary defines “inevitable” as “certain to happen, unavoidable”; which I interpret to mean that nothing we can do will change the destiny of the beta cells. If we apply this definition to beta cell secretory capacity, then the answer is clearly no: it’s not “inevitable”, since insulin sensitization by just about any means (e.g., weight loss, metformin, thiazolidinediones) can enhance secretory function and improve hyperglycemia or maintain normoglycemia for longer. However, if we apply this definition to long-term beta cell survival (which we should), then it becomes less clear. In experimental models of “glucolipotoxicity”, loss of beta cell mass is clearly reversible by ameliorating the metabolic insults, and is therefore not “inevitable”. But in humans, we simply cannot be sure as yet. The question boils down to whether our beta cells are just like those of rodents (not an unreasonable assumption, but by no means certain), and therefore potentially salvageable with sufficiently potent interventions, or whether they are fundamentally unique in some way, and therefore still prone to progressive and “unavoidable” loss of beta cell mass. Since no clinical intervention has yet been able to completely halt disease progression, are we actually changing the destiny of our beta cells? (If disease is still progressing slowly, isn’t it still just a matter of time?) Since we cannot simply biopsy the islets of large numbers of patients in a clinical trial, short of demonstrating that an intervention can actually sustain a long-term zero slope on the beta cell function vs. time curve, we cannot be certain that “inevitability” has truly been averted. Until we have definitive evidence that beta cell survival is truly sustained in humans, progressive long-term beta cell “failure” will still at least appear to be “inevitable”.

4/4/2010 - Question 2.  patiente 60 years old diabetes 2 dextro =4g/l bg despite one injection of 15u lantus usully receine 15u lantus a day at mid day what must i do

I’m not clear on the problem here. Can you clarify what you mean by “4g/l” (I’m assuming that’s a glucose reading of some sort)? Do you mean 4 mg/dL (!) or 4000 mg/L=400 mg/dL? What is the rest of the patient’s history? Concurrent medical problems and medications? Dietary and lifestyle patterns? Latest hemoglobin A1c reading? Glucose monitoring patterns?

12/18/2009 - Question 3.  Have a Diabetic on Lantus 80 u qpm, Avandamet 4/1000 BID. Pt doesn't want to do insulin after meals. I know that Byetta is not indicated with Insulin, but is it reasonable to try? Pt is overweight, so feel that Byetta may help with weight loss also. HbA1c was 11 before starting Lantus, but now down to 8.5. Also, how high can you go on Lantus? Is there an upper limit?, or is there a limit where you don't get much bang for the buck?

The use of exenatide added to insulin is a very interesting topic that is just beginning to be explored. It’s obviously off-label, as there are no proper trials to evaluate its efficacy, so I wouldn’t recommend it as a routine, but there are published reports of positive anecdotal experiences: check out the retrospective reports by Yoon et al. (Clin Therapeutics 2009; 31:1511), Sheffield et al. (Endocr Pract 2008; 14:285) or Viswanathan et al. (Endocr Pract 2007; 13:444). As one would expect from its complementary mechanism of action, exenatide does seem to provide additive glucose lowering, weight loss (which may very well mediate much of that glucose lowering), and often some measure of lowering insulin requirements, although that benefit seems to affect prandial insulin more than basal insulin (again, consistent with exenatide’s effects on glucose-dependent insulin secretion), so you may or may not notice a substantial reduction in the insulin requirements of your patient. Hypoglycemia is avoidable so long as the patient continues to monitor frequently and insulin doses are adjusted downwards in the event that control improves substantially. The one significant caution is the potential for GI side effects, mostly nausea, which can be limiting (and was responsible for many treatment discontinuations in the published reports). So if you feel you have no other viable options to bring the patient’s control down to target, it may be reasonable to try; start with the lower dose, and follow the patient’s readings frequently. As to your second question, one of the wonderful things about using insulin (any insulin – a unit is a unit) is that there is no true “ceiling” to the effective dose range, so long as you are guided by proper self-glucose monitoring. We have a number of patients in our clinic who require well over 200 units per day (sometimes even using concentrated U-500 insulin to limit the volume of the subcutaneous injection – check out Ballani et al. Diab Care 2006; 29:2504) and have achieved substantially improved control, whereas they previously failed other combination therapies. Yes, there is a diminishing return as you go higher up; once you’re above 100 units per day, it makes little sense to continue adjusting in only 2-4 unit increments; you’ll need to make bigger increments to get them to target, but there does not seem to be any kind of true plateau beyond which no further reduction occurs. (You do want to be sure that the high dose requirement is based on the patient’s inherent insulin resistance and not simply an out-of-control diet; chasing a patient’s erratic dietary habits with high-dose insulin will more than likely run you into significant hypoglycemia).

10/20/2009 - Question 4.  How do you match a particular insulin regimen with a given patient?

This is a complex question, and there is no single correct approach to optimize control in a given patient, let alone deal with the heterogeneity of different patients with different needs. The ultimate goal, of course, is to optimize glycemic control to meet recommended targets, but how that is achieved with insulin will depend on many factors, including the duration of the patient’s diabetes, their daily glycemic pattern, what other therapies have been beneficial but are not sufficient, each patient’s preferences, capabilities and lifestyle factors, as well as considerations of cost and availability/formulary concerns. There is good support in the literature that a simplified single-dose regimen of a basal insulin (i.e., NPH, glargine or detemir) given at bedtime can achieve adequate control in patients specifically with elevated fasting glucose with less weight gain and less hypoglycemia (as well as less inconvenience) than compared to more complex regimens, but it requires that the patient’s diabetes can still be controlled during the daytime with the patient’s current oral agents and dietary behaviors. A premixed insulin (e.g., 70/30 or similar mixture of intermediate/short-acting insulin) is another alternative that can provide daytime or 24-hour control, but is not as good as individualized mixing (such as traditional NPH-regular mixtures that are drawn up separately) to match the regimen to each patient’s unique daily glycemic pattern when it comes to intensification of control. Basal-bolus regimens consisting of a long-acting insulin plus a rapid-acting analogue (i.e., lispro, aspart or glulisine) is probably the most flexible option, and may provide superior control compared to other regimens, but it requires more injections and more frequent self-glucose monitoring, and should only be prescribed for patients who are sufficiently motivated and compliant to independently manage their own day-to-day glucose excursions. Even subcutaneous insulin pumps, if available, require frequent self-glucose monitoring and a motivated and knowledgeable patient, and should never be thought of as a fully automated insulin delivery system. The necessity of regular self-glucose monitoring that is appropriate for the complexity of each patient’s regimen, regardless of how the insulin is administered, must also be emphasized, and should factor into the decision regarding which regimen is ideal for a given patient. Referral for a specialized diabetes consultation should be considered if a patient wishes to explore the full spectrum of options that may be appropriate.

10/20/2009 - Question 5.  What is the overall nature of the concern with the use of the thiazolidinedione class of oral hypoglycemic agents, and rosiglitazone in particular?

As potent insulin sensitizers, thiazolidinediones (TZDs) are effective agents for controlling hyperglycemia in patients with type 2 diabetes, and they also improve many of the co-morbid conditions that often coexist with diabetes, such as hypertension, low levels of HDL-cholesterol, and inflammation, which collectively contribute to atherosclerotic risk. However, like many therapies currently used to manage type 2 diabetes, TZDs were not initially supported by any evidence of their effectiveness in reducing cardiovascular events, the major source of diabetes-related mortality. The two currently available TZDs, rosiglitazone and pioglitazone, have subtle differences in their actions on many of these surrogate treatment endpoints, so proper studies examining cardiovascular endpoints for both of these agents were important to help establish their overall risk-benefit profile. In 2005, the PROactive Study (Dormandy JA, Lancet 2005; 366:1279-89) demonstrated that pioglitazone effectively reduced one of the study’s secondary composite endpoints that included all-cause mortality and non-fatal MIs, although the study’s primary composite endpoint failed to reach statistical significance. Other pioglitazone studies examining surrogate endpoints such as carotid IMT (Mazzone T, JAMA 2006; 296:2572-81) and intravascular ultrasound (Nissen SE, JAMA 2008; 299:1561-73) have generally been consistent this clinical benefit. The current controversy principally surrounds rosiglitazone, which was the focus of a 2007 meta-analysis using non-adjudicated data from smaller, unpublished trials that suggested rosiglitazone may be associated with a paradoxical increase in myocardial infarctions (Nissen SE, N Eng J Med 2007; 356:2457-71). This is distinct from the known tendency of both TZDs to increase fluid retention and predispose to worsening CHF or peripheral edema. The definitive rosiglitazone study examining cardiovascular endpoints was the RECORD Study (Home PD, Lancet 2009; 373:2125-35), which did not detect a statistically significant increase in cardiovascular morbidity or mortality with rosiglitazone compared to non-TZD oral agents. However, this study also did not demonstrate any significant overall cardiovascular protection with rosiglitazone on any cardiovascular endpoint. Thus, there is reasonably good evidence that pioglitazone may be cardioprotective, while there is reasonably good evidence that rosiglitazone does not actually increase MI risk, but little evidence that it may be cardioprotective like pioglitazone. Given these subtle differences in the available evidence, there is generally no reason why rosiglitazone should be used if pioglitazone is equivalently available and tolerable to the patient. However, rosiglitazone remains an effective and viable option for glucose control if pioglitazone is unavailable or cannot be used.